US11449048B2 - Moving body control apparatus, moving body control method, and training method - Google Patents

Moving body control apparatus, moving body control method, and training method Download PDF

Info

Publication number
US11449048B2
US11449048B2 US16/362,846 US201916362846A US11449048B2 US 11449048 B2 US11449048 B2 US 11449048B2 US 201916362846 A US201916362846 A US 201916362846A US 11449048 B2 US11449048 B2 US 11449048B2
Authority
US
United States
Prior art keywords
moving body
machine learning
learning model
control command
control
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active, expires
Application number
US16/362,846
Other languages
English (en)
Other versions
US20190219998A1 (en
Inventor
Karthikk HARIHARA SUBRAMANIAN
Bin Zhou
Sheng Mei Shen
Sugiri Pranata Lim
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Panasonic Intellectual Property Corp of America
Original Assignee
Panasonic Intellectual Property Corp of America
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Panasonic Intellectual Property Corp of America filed Critical Panasonic Intellectual Property Corp of America
Priority to US16/362,846 priority Critical patent/US11449048B2/en
Publication of US20190219998A1 publication Critical patent/US20190219998A1/en
Assigned to PANASONIC INTELLECTUAL PROPERTY CORPORATION OF AMERICA reassignment PANASONIC INTELLECTUAL PROPERTY CORPORATION OF AMERICA ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: SHEN, SHENG MEI, ZHOU, BIN, HARIHARA SUBRAMANIAN, Karthikk, LIM, Sugiri Pranata
Application granted granted Critical
Publication of US11449048B2 publication Critical patent/US11449048B2/en
Active legal-status Critical Current
Adjusted expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/0011—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots associated with a remote control arrangement
    • G05D1/0016—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots associated with a remote control arrangement characterised by the operator's input device
    • G—PHYSICS
    • G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
    • G09G—ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION
    • G09G5/00—Control arrangements or circuits for visual indicators common to cathode-ray tube indicators and other visual indicators
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/0011—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots associated with a remote control arrangement
    • G05D1/0022—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots associated with a remote control arrangement characterised by the communication link
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/0011—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots associated with a remote control arrangement
    • G05D1/0033—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots associated with a remote control arrangement by having the operator tracking the vehicle either by direct line of sight or via one or more cameras located remotely from the vehicle
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/0011—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots associated with a remote control arrangement
    • G05D1/0038—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots associated with a remote control arrangement by providing the operator with simple or augmented images from one or more cameras located onboard the vehicle, e.g. tele-operation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Definitions

  • Patent Literature (PTL) 1 discloses a technique for controlling autonomous driving of a vehicle based on a pre-prepared three-dimensional map.
  • PTL 2 discloses a technique for controlling autonomous driving of a vehicle in accordance with a set driving route.
  • control of autonomous driving of a vehicle is limited in scope to a pre-prepared three-dimensional map.
  • control of autonomous driving of a vehicle is limited in scope to a set driving route.
  • one non-limiting and exemplary embodiment provides a moving body control apparatus, a moving body control method, and a training method capable of achieving autonomous driving of a moving body characterized by a higher degree of freedom than conventional art.
  • a moving body control apparatus for controlling a moving body and includes: an acquisition device for acquiring a control command for the moving body and an image of a view in a traveling direction of the moving body; and an information processing device that uses a machine learning model to output a control parameter for controlling the moving body, using the control command and the image acquired by the acquisition device as inputs.
  • a moving body control method for controlling a moving body, and includes: a first step of acquiring a control command for the moving body and an image of a view in a traveling direction of the moving body; and a second step of using a machine learning model to output a control parameter for controlling the moving body, using the control command and the image acquired in the first step as inputs.
  • a training method is for a moving body control apparatus that controls a moving body using a machine learning model, and includes a first step of training the machine learning model using first training data that takes a control command for the moving body and an image of a view in a traveling direction of the moving body as inputs, and includes a control parameter for controlling the moving body as a correct answer.
  • a moving body control apparatus, a moving body control method, and a training method according to one aspect of the present disclosure are capable of achieving autonomous driving characterized by a higher degree of freedom than conventional art.
  • FIG. 1 is a block diagram of a configuration of a moving body control apparatus according to an embodiment.
  • FIG. 2 schematically illustrates one example in which a moving body control apparatus according to an embodiment is installed in a moving body, which is control target.
  • FIG. 3 schematically illustrates a configuration of a machine learning model according to an embodiment.
  • FIG. 4 is a flowchart of a training method for training a machine learning model according to an embodiment.
  • FIG. 5A schematically illustrates training of a machine learning model according to an embodiment using second training data.
  • FIG. 5B schematically illustrates training of a machine learning model according to an embodiment using general image training data.
  • FIG. 6 schematically illustrates training of a machine learning model according to an embodiment using first training data.
  • FIG. 7 illustrates a result of a test in which a machine learning model according to an embodiment was trained using constraint conditions.
  • FIG. 8 is a flowchart of moving body control process according to an embodiment.
  • a moving body control apparatus for controlling a moving body and includes: an acquisition device for acquiring a control command for the moving body and an image of a view in a traveling direction of the moving body; and an information processing device that uses a machine learning model to output a control parameter for controlling the moving body, using the control command and the image acquired by the acquisition device as inputs.
  • control command is a signal indicating an intention to maneuver the moving body, such as turn left, turn right, stop, go forward, make a U-turn, change lanes, etc.
  • control parameter is a signal indicating a physical quantity for controlling the moving body 100 , such as a speed and/or a steering angle.
  • the moving body control apparatus described above controls the moving body using a control command for the moving body and an image of a view in the traveling direction of the moving body.
  • a moving body control method is for controlling a moving body, and includes: a first step of acquiring a control command for the moving body and an image of a view in a traveling direction of the moving body; and a second step of using a machine learning model to output a control parameter for controlling the moving body, using the control command and the image acquired in the first step as inputs.
  • the moving body control method described above controls the moving body using a control command for the moving body and an image of a view in the traveling direction of the moving body.
  • a training method is for a moving body control apparatus that controls a moving body using a machine learning model, and includes a first step of training the machine learning model using first training data that takes a control command for the moving body and an image of a view in a traveling direction of the moving body as inputs, and includes a control parameter for controlling the moving body as a correct answer.
  • the machine learning model trained via the training method described above controls the moving body using a control command for the moving body and an image of a view in the traveling direction of the moving body.
  • FIG. 1 is a block diagram of a configuration of a moving body control apparatus 1 according to the embodiment.
  • FIG. 2 schematically illustrates one example in which the moving body control apparatus 1 is installed in a moving body 100 , which is control target.
  • the moving body 100 is illustrated as an automobile, the control target of the moving body control apparatus 1 is not limited to an automobile; the control target may be any controllable moving body, such as an autonomous robot, for example.
  • the control target may be any controllable moving body, such as an autonomous robot, for example.
  • some or all of the elements in the moving body control apparatus 1 need not be equipped on the moving body 100 , which is the control target.
  • the moving body control apparatus 1 includes an acquisition device 10 and an information processing device 20 .
  • the acquisition device 10 acquires a control command for the moving body 100 and an image of a view in the traveling direction of the moving body 100 .
  • the control command is a signal indicating an intention to maneuver moving body 100 , such as turn left, turn right, stop, go forward, make a U-turn, change lanes, etc.
  • the control command is exemplified as including at least one of turn left, turn right, stop, go forward, make a U-turn, and change lanes.
  • the acquisition device 10 may include, for example, an input device that acquires a control command for the moving body 100 by receiving, from a user of the moving body control apparatus 1 , an input of a control command for the moving body 100 .
  • the acquisition device 10 may include, for example, a receiver that acquires a control command for the moving body 100 by receiving a control command for the moving body 100 output from an external device with which communication is possible.
  • the acquisition device 10 may include an image capturing device that is disposed in a front portion of the moving body 100 and acquires an image of a view in the traveling direction of the moving body 100 by capturing an image of a view in the traveling direction of the moving body 100 .
  • the image capturing device may be a digital video camera including an image sensor.
  • the acquisition device 10 acquires each frame of a video captured by the digital video camera as an image of a view in the traveling direction of the moving body 100 .
  • the acquisition device 10 may include, for example, a receiver that acquires an image of a view in the traveling direction of the moving body 100 by receiving, from an external image capturing device with which communication is possible, an image of a view in the traveling direction of the moving body 100 captured by the image capturing device.
  • the information processing device 20 uses a machine learning model to output a control parameter for controlling the moving body 100 , using the control command and the image acquired by the acquisition device 10 as inputs.
  • the control parameter is a signal indicating a physical quantity for controlling the moving body 100 , such as a speed and/or a steering angle.
  • the control parameter is exemplified as including at least one of a speed and a steering angle.
  • the information processing device 20 trains a machine learning model by using training data or a constraint condition.
  • the information processing device 20 may include, for example, a processor capable of communicating with the acquisition device 10 and memory. In such cases, the information processing device 20 can implement various processes by the processor executing a program stored in the memory.
  • the machine learning model is a model that is trained to output a control parameter when a control command and an image of a view in the traveling direction of the moving body 100 is input.
  • the machine learning model may be, for example, a neural network model.
  • the machine learning model is exemplified as a neural network model.
  • FIG. 3 schematically illustrates a configuration of a machine learning model 30 used by the information processing device 20 .
  • the machine learning model 30 includes a first machine learning model 40 and a second machine learning model 50 .
  • the first machine learning model 40 is a model that is trained to, when an image of a view in the traveling direction of the moving body 100 is input, calculate a feature amount of an object from the input image.
  • the second machine learning model 50 is a model that is trained to calculate a control parameter when the feature amount output from the first machine learning model 40 and a control command are input.
  • the second machine learning model 50 includes a first control command machine learning model 51 A and a second control command machine learning model 51 B through Nth (N is an integer greater than or equal to 2) control command machine learning model 51 N respectively corresponding to N control commands input into the machine learning model 30 .
  • N is an integer greater than or equal to 2
  • control command machine learning model 51 N respectively corresponding to N control commands input into the machine learning model 30 .
  • control command learning models 51 are referred to as control command learning models 51 .
  • Each control command learning model 51 is a model that is trained to calculate a control parameter when a feature amount associated the control command that the model correspond to is input.
  • the second machine learning model 50 selects a control command learning model 51 that corresponds to the input control command.
  • the second machine learning model 50 outputs a control parameter using, as an input, a feature amount that is output from the first machine learning model 40 and associated with the control command.
  • FIG. 4 is a flowchart of a training method used by the information processing device 20 for training the machine learning model 30 .
  • the information processing device 20 trains the first machine learning model 40 as a classifier, using training data that takes images as an input and includes labels attached to the images as correct answers (step S 10 ).
  • Applicable training data includes training data that takes images of a view in the traveling direction of the moving body 100 as an input, and includes control commands as correct answers (hereinafter referred to as “second training data”), and training data that takes general images not necessarily related to the moving body 100 as an input, and includes label information attached to the general images as correct answers (hereinafter referred to as “general image training data”).
  • the training may use both the second training data and the general image training data.
  • step S 10 for example, the second training data is used.
  • the training is, for example, performed in steps, using the first the general image training data first and the second training data subsequently.
  • sets of an image of a view in the traveling direction of the moving body 100 and a control command used in the second training data are acquired in advance by way of the user of the moving body control apparatus 1 operating the moving body 100 .
  • the user of the moving body control apparatus 1 repeatedly operates the moving body 100 in accordance with the N control commands to be inputted into the machine learning model 30 , to acquire a plurality of images of a view in the traveling direction of the moving body 100 , each corresponding to one of the N control commands.
  • the user of the moving body control apparatus 1 then groups the N control commands and the corresponding plurality of images of a view in the traveling direction of the moving body 100 to generate the second training data.
  • FIG. 5A schematically illustrates the information processing device 20 in step S 10 training the first machine learning model 40 as a classifier using the second training data.
  • FIG. 5B schematically illustrates the information processing device 20 in step S 10 training the first machine learning model 40 as a classifier using the general image training data.
  • the information processing device 20 inputs, into the first machine learning model 40 , a plurality of images of a view in the traveling direction of the moving body 100 labeled with control commands as correct answers, for the N control commands.
  • the information processing device 20 trains the first machine learning model 40 as a classifier that classifies each input image of a view in the traveling direction of the moving body 100 into the corresponding control command.
  • the information processing device 20 trains the machine learning model 30 , via a regression technique, using training data that takes control commands and images of a view in the traveling direction of the moving body 100 as inputs, and includes control parameters as correct answers (hereinafter referred to as “first training data”) (step S 20 ).
  • sets of a control command, an image of a view in the traveling direction of the moving body 100 , and a control parameter used in the first training data are acquired in advance by way of the user of the moving body control apparatus 1 operating the moving body 100 .
  • the user of the moving body control apparatus 1 repeatedly operates the moving body 100 in accordance with the N control commands to be inputted into the machine learning model 30 , to acquire a plurality of images of a view in the traveling direction of the moving body 100 , each corresponding to one of the N control commands, and a plurality of control parameters, each corresponding to one of the N control commands.
  • the user of the moving body control apparatus 1 then groups the plurality of images of a view in the traveling direction of the moving body 100 , the corresponding control commands, and the corresponding control parameter to generate the first training data.
  • FIG. 6 schematically illustrates the information processing device 20 in step S 20 training the machine learning model 30 , via a regression technique, using the first training data.
  • the information processing device 20 trains the machine learning model 30 , via a regression technique, about a plurality of images of a view in the traveling direction of the moving body 100 , by using the plurality of images of a view in the traveling direction of the moving body 100 and corresponding control commands as inputs, and using corresponding control parameters as correct answers.
  • the machine learning model 30 is trained to a level that is usable by the information processing device 20 to control the moving body 100 (hereinafter referred to as a “usable level”).
  • the information processing device 20 may train the machine learning model 30 via a regression technique, using the first training data, in a state in which the first machine learning model 40 is fixed as a “trained” model (i.e., in a state in which the first machine learning model 40 is no longer to be further trained).
  • step S 30 the information processing device 20 trains the machine learning model 30 via a regression technique, using a constraint condition that restricts an operating state of the moving body 100 (step S 30 ).
  • the processing in step S 30 trains the moving body 100 to move more smoothly.
  • smooth movement of the moving body 100 is movement unaccompanied by sharp changes in moving velocity and sharp changes in steering angle.
  • a constraint condition may be, for example, a condition that imposes an upper limit for a rate of change in the moving velocity v of the moving body 100 per unit time, expressed as dv/dt (in other words, the rate of acceleration of the moving body 100 ) (hereinafter, this condition is referred to as “constraint condition 1”).
  • the constraint condition may be, for example, a condition stipulating that when the moving body 100 is driving on a road marked with a white line, the moving body 100 is not to cross the white line (hereinafter, this condition is referred to as “constraint condition 2”).
  • the constraint condition may be, for example, a condition that imposes an upper limit for a rate of change in the steering angle of the moving body 100 per unit time or per unit distance.
  • FIG. 7 illustrates a result of a test in which the information processing device 20 trained the machine learning model 30 via a regression technique, using constraint condition 1 and constraint condition 2.
  • smoothness 71 is a benchmark indicating a degree of smoothness in the movement of the moving body 100 reflecting the rate of change in moving velocity of the moving body 100 per unit time and the rate of change in the steering angle of the moving body 100 per unit time. The smaller the value is, the smoother the movement of the moving body 100 is.
  • white-line-crossing count 72 indicates the number of times the moving body 100 crossed a white line when driving on a road marked with a white line.
  • the machine learning model 30 is capable of controlling moving body 100 such that the moving body 100 moves more smoothly and does not cross a white line.
  • the information processing device 20 may train the machine learning model 30 via a regression technique, using a constraint condition, in a state in which the first machine learning model 40 is fixed as a “trained” model.
  • the information processing device 20 trains the machine learning model 30 via a regression technique, using a traffic rule to be obeyed by the moving body 100 (step S 40 ).
  • the processing in step S 40 is training for causing the moving body 100 move in accordance with a traffic rule.
  • a traffic rule may be, for example, a condition stipulating that traffic lights are to be obeyed.
  • a traffic rule may be, for example, a condition stipulating that driving lanes are not to be changed.
  • the information processing device 20 may train the machine learning model 30 via a regression technique, using a traffic rule, in a state in which the first machine learning model 40 is fixed as a “trained” model.
  • step S 40 the information processing device 20 ends training of the machine learning model 30 .
  • the information processing device 20 is described as performing the processing in step S 40 after the processing in step S 30 is complete. However, in another example of the training method, the information processing device 20 may perform the processing in step S 30 after the processing in step S 40 is complete.
  • both the processing in step S 30 and the processing in step S 40 are for improving the quality of movement of the moving body 100 . Accordingly, depending on the desired level of quality of movement of the moving body 100 , the information processing device 20 may perform only one of the processing in step S 30 and the processing in step S 40 , and may perform neither the processing in step S 30 nor the processing in step S 40 .
  • the information processing device 20 can train the machine learning model 30 to a usable level.
  • a training method in which the processing in step S 25 is performed instead of the processing in step S 10 and the processing in step S 20 reduced the learning efficiency of machine learning model 30 compared to the training method in which the processing in step S 10 and the processing in step S 20 were performed. Accordingly, the training method in which the processing in step S 10 and the processing in step S 20 are performed is beneficial.
  • step S 25 is processing in which the information processing device 20 trains the entire machine learning model 30 via a regression technique, using the first training data, in a state in which the first machine learning model 40 has not been trained in advance (i.e., in a state in which step S 10 has not been performed).
  • the moving body control apparatus 1 performs a moving body control process for controlling the moving body 100 by using the machine learning model 30 trained via the above-described training method.
  • FIG. 8 is a flowchart of the moving body control process.
  • the moving body control process starts when the user of the moving body control apparatus 1 performs an operation on the moving body control apparatus 1 signifying starting the moving body control process.
  • the acquisition device 10 When the moving body control process starts, the acquisition device 10 successively acquires images of a view in the traveling direction of the moving body 100 (step S 100 ).
  • step S 110 when the acquisition device 10 acquires a control command for the moving body 100 (yes in step S 110 after repeated results of no in step S 110 ), the information processing device 20 inputs, into the machine learning model 30 , the control command acquired by the acquisition device 10 and the most recent image of a view in the traveling direction of the moving body 100 acquired by the acquisition device 10 (step S 120 ).
  • the first machine learning model 40 calculates, from the image, a feature amount associated with the control command and indicating a feature of an object in the image (step S 130 ).
  • the second machine learning model 50 selects the control command learning model 51 that corresponds to the control command (step S 140 ). Then, the selected control command learning model 51 calculates a control parameter from the feature amount associated with the control command (step S 150 ).
  • the information processing device 20 updates the control parameter such that the difference between the current control parameter and the calculated control parameter decreases (step S 160 ), and outputs the updated control parameter.
  • the method for determining the update interval may be, for example, Proportional Integral Differential (PID) control. Note that the control parameter may be updated with the calculated control parameter on the spot, without using a gradual change.
  • the information processing device 20 determines whether the current control parameter and the calculated control parameter are equal or not (step S 170 ).
  • step S 170 when the current control parameter and the calculated control parameter are determined to be equal (yes in step S 170 ), the moving body control apparatus 1 returns to step S 110 , and performs the subsequent steps.
  • step S 170 when the current control parameter and the calculated control parameter are determined to not be equal (no in step S 170 ), the moving body control apparatus 1 returns to step S 160 , and performs the subsequent steps.
  • the moving body control apparatus 1 having the above configuration controls the moving body 100 using a control command for the moving body 100 and an image of a view in the traveling direction of the moving body 100 .
  • the moving body 100 is autonomously driven using the moving body control apparatus 1 , there is no need for a pre-prepared three-dimensional map, and there is no need to specify a specific driving position. Moreover, selection of something other than a fixed driving route is possible.
  • the moving body control apparatus 1 it is possible to autonomously drive the moving body 100 with a higher degree of freedom than conventional art.
  • the information processing device 20 may include a plurality of computer devices capable of communicating with one another, and the various processes performed by the information processing device 20 may be implemented by distributed computing or cloud computing.
  • some or all of the elements included in the information processing device 20 may be realized via a single system large scale integrated (LSI) circuit.
  • LSI large scale integrated
  • a system LSI circuit is a multifunctional LSI circuit manufactured by integrating a plurality of units on a single chip, and is specifically a computer system including, for example, a microprocessor, ROM (Read Only Memory), and RAM (Random Access Memory).
  • a computer program is stored in the ROM.
  • the system LSI circuit achieves its function as a result of the microprocessor operating according to the computer program.
  • system LSI circuit is used, but depending on the degree of integration, the circuit may also referred to as IC, LSI circuit, super LSI circuit, or ultra LSI circuit. Moreover, the method of circuit integration is not limited to LSI. Integration may be realized with a specialized circuit or a general purpose processor. After the LSI circuit is manufactured, a field programmable gate array (FPGA) or a reconfigurable processor capable of reconfiguring the connections and settings of the circuit cells in the LSI circuit may be used.
  • FPGA field programmable gate array
  • Each of the elements in each of the above embodiments may be configured in the form of an exclusive hardware product, or may be realized by executing a software program suitable for the element.
  • Each of the elements may be realized by means of a program executing unit, such as a CPU or a processor, reading and executing the software program recorded on a recording medium such as a hard disk or semiconductor memory.
  • the present disclosure is widely applicable in control devices for control a moving body.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Automation & Control Theory (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Computer Hardware Design (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Mechanical Engineering (AREA)
  • Robotics (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
  • Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
  • Image Analysis (AREA)
US16/362,846 2017-06-28 2019-03-25 Moving body control apparatus, moving body control method, and training method Active 2038-10-10 US11449048B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/362,846 US11449048B2 (en) 2017-06-28 2019-03-25 Moving body control apparatus, moving body control method, and training method

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US201762525979P 2017-06-28 2017-06-28
PCT/US2018/038433 WO2019005547A1 (en) 2017-06-28 2018-06-20 MOBILE BODY CONTROL APPARATUS, MOBILE BODY CONTROL METHOD, AND LEARNING METHOD
US16/362,846 US11449048B2 (en) 2017-06-28 2019-03-25 Moving body control apparatus, moving body control method, and training method

Related Parent Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2018/038433 Continuation WO2019005547A1 (en) 2017-06-28 2018-06-20 MOBILE BODY CONTROL APPARATUS, MOBILE BODY CONTROL METHOD, AND LEARNING METHOD

Publications (2)

Publication Number Publication Date
US20190219998A1 US20190219998A1 (en) 2019-07-18
US11449048B2 true US11449048B2 (en) 2022-09-20

Family

ID=64742591

Family Applications (1)

Application Number Title Priority Date Filing Date
US16/362,846 Active 2038-10-10 US11449048B2 (en) 2017-06-28 2019-03-25 Moving body control apparatus, moving body control method, and training method

Country Status (3)

Country Link
US (1) US11449048B2 (ja)
JP (1) JP7050006B2 (ja)
WO (1) WO2019005547A1 (ja)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220185296A1 (en) * 2017-12-18 2022-06-16 Plusai, Inc. Method and system for human-like driving lane planning in autonomous driving vehicles
US11643086B2 (en) 2017-12-18 2023-05-09 Plusai, Inc. Method and system for human-like vehicle control prediction in autonomous driving vehicles
US11650586B2 (en) 2017-12-18 2023-05-16 Plusai, Inc. Method and system for adaptive motion planning based on passenger reaction to vehicle motion in autonomous driving vehicles

Families Citing this family (24)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018176000A1 (en) 2017-03-23 2018-09-27 DeepScale, Inc. Data synthesis for autonomous control systems
JP7050006B2 (ja) * 2017-06-28 2022-04-07 パナソニック インテレクチュアル プロパティ コーポレーション オブ アメリカ 移動体制御装置、移動体制御方法、及び学習方法
US11893393B2 (en) 2017-07-24 2024-02-06 Tesla, Inc. Computational array microprocessor system with hardware arbiter managing memory requests
US11157441B2 (en) 2017-07-24 2021-10-26 Tesla, Inc. Computational array microprocessor system using non-consecutive data formatting
US10671349B2 (en) 2017-07-24 2020-06-02 Tesla, Inc. Accelerated mathematical engine
US11409692B2 (en) 2017-07-24 2022-08-09 Tesla, Inc. Vector computational unit
US11561791B2 (en) 2018-02-01 2023-01-24 Tesla, Inc. Vector computational unit receiving data elements in parallel from a last row of a computational array
US11215999B2 (en) 2018-06-20 2022-01-04 Tesla, Inc. Data pipeline and deep learning system for autonomous driving
US11361457B2 (en) 2018-07-20 2022-06-14 Tesla, Inc. Annotation cross-labeling for autonomous control systems
US11636333B2 (en) 2018-07-26 2023-04-25 Tesla, Inc. Optimizing neural network structures for embedded systems
US11562231B2 (en) 2018-09-03 2023-01-24 Tesla, Inc. Neural networks for embedded devices
WO2020077117A1 (en) 2018-10-11 2020-04-16 Tesla, Inc. Systems and methods for training machine models with augmented data
US11196678B2 (en) 2018-10-25 2021-12-07 Tesla, Inc. QOS manager for system on a chip communications
US11816585B2 (en) 2018-12-03 2023-11-14 Tesla, Inc. Machine learning models operating at different frequencies for autonomous vehicles
US11537811B2 (en) 2018-12-04 2022-12-27 Tesla, Inc. Enhanced object detection for autonomous vehicles based on field view
US11610117B2 (en) 2018-12-27 2023-03-21 Tesla, Inc. System and method for adapting a neural network model on a hardware platform
US11150664B2 (en) 2019-02-01 2021-10-19 Tesla, Inc. Predicting three-dimensional features for autonomous driving
US10997461B2 (en) 2019-02-01 2021-05-04 Tesla, Inc. Generating ground truth for machine learning from time series elements
US11567514B2 (en) 2019-02-11 2023-01-31 Tesla, Inc. Autonomous and user controlled vehicle summon to a target
US10956755B2 (en) 2019-02-19 2021-03-23 Tesla, Inc. Estimating object properties using visual image data
IT201900002853A1 (it) * 2019-02-27 2020-08-27 Magneti Marelli Spa "Sistema per ottenere la predizione di un’azione di un veicolo e procedimento corrispondente"
DE112020006315B4 (de) * 2020-02-27 2023-12-14 Mitsubishi Electric Corporation Robotersteuervorrichtung, robotersteuerverfahren und vorrichtung zur erzeugung von lernmodellen
US11203361B2 (en) * 2020-05-05 2021-12-21 StradVision, Inc. Method for performing on-device learning of machine learning network on autonomous vehicle by using multi-stage learning with adaptive hyper-parameter sets and device using the same
JP7466403B2 (ja) * 2020-08-03 2024-04-12 キヤノン株式会社 制御装置、リソグラフィー装置、制御方法および物品製造方法

Citations (26)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH11282530A (ja) 1998-03-27 1999-10-15 Sanyo Electric Co Ltd 車両運行システム
JP2009199572A (ja) 2008-01-25 2009-09-03 Kazuo Iwane 三次元機械地図、三次元機械地図生成装置、ナビゲーション装置及び自動運転装置
US7783391B2 (en) * 2005-10-28 2010-08-24 Electronics And Telecommunications Research Institute Apparatus and method for controlling vehicle by teeth-clenching
JP2011238054A (ja) 2010-05-11 2011-11-24 Toyota Motor Corp 走行軌道作成装置
US20120166200A1 (en) * 2010-12-23 2012-06-28 Electronics And Telecommunications Research Institute System and method for integrating gesture and sound for controlling device
JP2013113274A (ja) 2011-11-30 2013-06-10 Toyota Motor Corp エンジン制御装置
US20150032258A1 (en) * 2013-07-29 2015-01-29 Brain Corporation Apparatus and methods for controlling of robotic devices
US20150217449A1 (en) * 2014-02-03 2015-08-06 Brain Corporation Apparatus and methods for control of robot actions based on corrective user inputs
CN106080590A (zh) 2016-06-12 2016-11-09 百度在线网络技术（北京）有限公司 车辆控制方法和装置以及决策模型的获取方法和装置
US20160334229A1 (en) * 2015-05-13 2016-11-17 Uber Technologies, Inc. Autonomous vehicle operated with guide assistance of human driven vehicles
US20170213457A1 (en) * 2014-07-23 2017-07-27 Hatsumeiya Co, Ltd Automobile and computing system
US20180032082A1 (en) * 2016-01-05 2018-02-01 Mobileye Vision Technologies Ltd. Machine learning navigational engine with imposed constraints
US20180247160A1 (en) * 2017-02-27 2018-08-30 Mohsen Rohani Planning system and method for controlling operation of an autonomous vehicle to navigate a planned path
US20190034818A1 (en) * 2017-07-28 2019-01-31 Fanuc Corporation Machine learning device and user identification device
US20190118394A1 (en) * 2017-10-19 2019-04-25 Canon Kabushiki Kaisha Control apparatus, robot system, method for operating control apparatus, and storage medium
US20190188542A1 (en) * 2017-12-18 2019-06-20 Ford Global Technologies, Llc Using Deep Video Frame Prediction For Training A Controller Of An Autonomous Vehicle
US20190219998A1 (en) * 2017-06-28 2019-07-18 Panasonic Intellectual Property Corporation Of America Moving body control apparatus, moving body control method, and training method
US20190294159A1 (en) * 2016-11-30 2019-09-26 Nissan North America, Inc. Autonomous Vehicle Remote Support Mapping Interface
US20190310648A1 (en) * 2018-04-09 2019-10-10 Toyota Jidosha Kabushiki Kaisha Machine Learning Enhanced Vehicle Merging
US20190317502A1 (en) * 2016-12-01 2019-10-17 SZ DJI Technology Co., Ltd. Method, apparatus, device, and system for controlling unmanned aerial vehicle
US20190391582A1 (en) * 2019-08-20 2019-12-26 Lg Electronics Inc. Apparatus and method for controlling the driving of a vehicle
US10576636B1 (en) * 2019-04-12 2020-03-03 Mujin, Inc. Method and control system for and updating camera calibration for robot control
US10599155B1 (en) * 2014-05-20 2020-03-24 State Farm Mutual Automobile Insurance Company Autonomous vehicle operation feature monitoring and evaluation of effectiveness
US20200242154A1 (en) * 2019-01-24 2020-07-30 Olympus Corporation Image file generating device and image file generating method
US20210012193A1 (en) * 2019-07-11 2021-01-14 Fujitsu Limited Machine learning method and machine learning device
US10922585B2 (en) * 2018-03-13 2021-02-16 Recogni Inc. Deterministic labeled data generation and artificial intelligence training pipeline

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5531612B2 (ja) * 2009-12-25 2014-06-25 ソニー株式会社 情報処理装置、情報処理方法、プログラム、制御対象機器および情報処理システム
BR112012017199A2 (pt) * 2009-12-25 2018-07-31 Ricoh Co Ltd aparelho para identificação de objetos, aparelho para controle de corpos em movimento e aparelho para fornecer informações
JP5790677B2 (ja) * 2013-02-15 2015-10-07 トヨタ自動車株式会社 移動制御装置、移動体制御方法、及び制御プログラム
JP6020328B2 (ja) * 2013-04-18 2016-11-02 トヨタ自動車株式会社 移動体制御装置、移動体制御方法、及び制御プログラム

Patent Citations (27)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH11282530A (ja) 1998-03-27 1999-10-15 Sanyo Electric Co Ltd 車両運行システム
US7783391B2 (en) * 2005-10-28 2010-08-24 Electronics And Telecommunications Research Institute Apparatus and method for controlling vehicle by teeth-clenching
JP2009199572A (ja) 2008-01-25 2009-09-03 Kazuo Iwane 三次元機械地図、三次元機械地図生成装置、ナビゲーション装置及び自動運転装置
JP2011238054A (ja) 2010-05-11 2011-11-24 Toyota Motor Corp 走行軌道作成装置
US20120166200A1 (en) * 2010-12-23 2012-06-28 Electronics And Telecommunications Research Institute System and method for integrating gesture and sound for controlling device
JP2013113274A (ja) 2011-11-30 2013-06-10 Toyota Motor Corp エンジン制御装置
US20150032258A1 (en) * 2013-07-29 2015-01-29 Brain Corporation Apparatus and methods for controlling of robotic devices
US20150217449A1 (en) * 2014-02-03 2015-08-06 Brain Corporation Apparatus and methods for control of robot actions based on corrective user inputs
US10599155B1 (en) * 2014-05-20 2020-03-24 State Farm Mutual Automobile Insurance Company Autonomous vehicle operation feature monitoring and evaluation of effectiveness
US20170213457A1 (en) * 2014-07-23 2017-07-27 Hatsumeiya Co, Ltd Automobile and computing system
US20160334229A1 (en) * 2015-05-13 2016-11-17 Uber Technologies, Inc. Autonomous vehicle operated with guide assistance of human driven vehicles
US20180032082A1 (en) * 2016-01-05 2018-02-01 Mobileye Vision Technologies Ltd. Machine learning navigational engine with imposed constraints
US20170357257A1 (en) 2016-06-12 2017-12-14 Baidu Online Network Technology (Beijing) Co., Ltd. Vehicle control method and apparatus and method and apparatus for acquiring decision-making model
CN106080590A (zh) 2016-06-12 2016-11-09 百度在线网络技术（北京）有限公司 车辆控制方法和装置以及决策模型的获取方法和装置
US20190294159A1 (en) * 2016-11-30 2019-09-26 Nissan North America, Inc. Autonomous Vehicle Remote Support Mapping Interface
US20190317502A1 (en) * 2016-12-01 2019-10-17 SZ DJI Technology Co., Ltd. Method, apparatus, device, and system for controlling unmanned aerial vehicle
US20180247160A1 (en) * 2017-02-27 2018-08-30 Mohsen Rohani Planning system and method for controlling operation of an autonomous vehicle to navigate a planned path
US20190219998A1 (en) * 2017-06-28 2019-07-18 Panasonic Intellectual Property Corporation Of America Moving body control apparatus, moving body control method, and training method
US20190034818A1 (en) * 2017-07-28 2019-01-31 Fanuc Corporation Machine learning device and user identification device
US20190118394A1 (en) * 2017-10-19 2019-04-25 Canon Kabushiki Kaisha Control apparatus, robot system, method for operating control apparatus, and storage medium
US20190188542A1 (en) * 2017-12-18 2019-06-20 Ford Global Technologies, Llc Using Deep Video Frame Prediction For Training A Controller Of An Autonomous Vehicle
US10922585B2 (en) * 2018-03-13 2021-02-16 Recogni Inc. Deterministic labeled data generation and artificial intelligence training pipeline
US20190310648A1 (en) * 2018-04-09 2019-10-10 Toyota Jidosha Kabushiki Kaisha Machine Learning Enhanced Vehicle Merging
US20200242154A1 (en) * 2019-01-24 2020-07-30 Olympus Corporation Image file generating device and image file generating method
US10576636B1 (en) * 2019-04-12 2020-03-03 Mujin, Inc. Method and control system for and updating camera calibration for robot control
US20210012193A1 (en) * 2019-07-11 2021-01-14 Fujitsu Limited Machine learning method and machine learning device
US20190391582A1 (en) * 2019-08-20 2019-12-26 Lg Electronics Inc. Apparatus and method for controlling the driving of a vehicle

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
International Search Report dated Sep. 11, 2018 in International (PCT) Application PCT/US2018/038433.

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220185296A1 (en) * 2017-12-18 2022-06-16 Plusai, Inc. Method and system for human-like driving lane planning in autonomous driving vehicles
US20220185295A1 (en) * 2017-12-18 2022-06-16 Plusai, Inc. Method and system for personalized driving lane planning in autonomous driving vehicles
US11643086B2 (en) 2017-12-18 2023-05-09 Plusai, Inc. Method and system for human-like vehicle control prediction in autonomous driving vehicles
US11650586B2 (en) 2017-12-18 2023-05-16 Plusai, Inc. Method and system for adaptive motion planning based on passenger reaction to vehicle motion in autonomous driving vehicles
US12060066B2 (en) * 2017-12-18 2024-08-13 Plusai, Inc. Method and system for human-like driving lane planning in autonomous driving vehicles
US12071142B2 (en) * 2017-12-18 2024-08-27 Plusai, Inc. Method and system for personalized driving lane planning in autonomous driving vehicles

Also Published As

Publication number Publication date
US20190219998A1 (en) 2019-07-18
JP7050006B2 (ja) 2022-04-07
WO2019005547A1 (en) 2019-01-03
JP2020525331A (ja) 2020-08-27

Similar Documents

Publication Publication Date Title
US11449048B2 (en) Moving body control apparatus, moving body control method, and training method
EP3825803B1 (en) Collision-avoidance system for autonomous-capable vehicles
JP7086911B2 (ja) 自動運転車両のためのリアルタイム意思決定
CN111868641B (zh) 用于产生用于训练车辆控制设备的人工智能模块的训练数据组的方法
EP3588226B1 (en) Method and arrangement for generating control commands for an autonomous road vehicle
US11125572B2 (en) Electronic control device for autonomous driving of a vehicle
US20200218280A1 (en) Autonomous driving device
CN110843789B (zh) 一种基于时序卷积网络的车辆换道意图预测方法
EP2960131A2 (en) Warning device and travel control device
CN112840350A (zh) 自动驾驶车辆规划和预测
WO2022132586A1 (en) Lane change gap finder
JP7216766B2 (ja) 車両制御装置
US20190164051A1 (en) Systems and methods for training and controlling an artificial neural network with discrete vehicle driving commands
US12002353B2 (en) Method and system for providing environmental data
US12008743B2 (en) Hazard detection ensemble architecture system and method
DE102021203588A1 (de) Verfahren und Steuergerät zum Schätzen eines Verhaltens eines Systems
CN108528457B (zh) 运动规划的方法、设备和具有指令的计算机可读存储介质
KR102376615B1 (ko) 주행 로봇의 제어 방법 및 그 장치
DE102020201931A1 (de) Verfahren zum Trainieren wenigstens eines Algorithmus für ein Steuergerät eines Kraftfahrzeugs, Verfahren zur Optimierung eines Verkehrsflusses in einer Region, Computerprogrammprodukt sowie Kraftfahrzeug
JP2018103859A (ja) 操舵支援方法及び操舵支援装置
JP7268321B2 (ja) 駐車支援方法及び駐車支援装置
Shah et al. Comparing Traditional Computer Vision Algorithms and Deep Convolutional Neural Networks as Self Driving Algorithms for Use in Dynamic Conditions
CN112693470A (zh) 一种避免车辆转弯风险的方法和装置
CN113327456A (zh) 一种车道结构检测方法及装置
EP4245629A1 (en) Method and computer system for controlling the movement of a host vehicle

Legal Events

Date Code Title Description
FEPP Fee payment procedure

Free format text: ENTITY STATUS SET TO UNDISCOUNTED (ORIGINAL EVENT CODE: BIG.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

STPP Information on status: patent application and granting procedure in general

Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION

AS Assignment

Owner name: PANASONIC INTELLECTUAL PROPERTY CORPORATION OF AMERICA, CALIFORNIA

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:HARIHARA SUBRAMANIAN, KARTHIKK;ZHOU, BIN;SHEN, SHENG MEI;AND OTHERS;SIGNING DATES FROM 20190214 TO 20190219;REEL/FRAME:050187/0990

Owner name: PANASONIC INTELLECTUAL PROPERTY CORPORATION OF AME

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:HARIHARA SUBRAMANIAN, KARTHIKK;ZHOU, BIN;SHEN, SHENG MEI;AND OTHERS;SIGNING DATES FROM 20190214 TO 20190219;REEL/FRAME:050187/0990

STPP Information on status: patent application and granting procedure in general

Free format text: NON FINAL ACTION MAILED

STPP Information on status: patent application and granting procedure in general

Free format text: RESPONSE TO NON-FINAL OFFICE ACTION ENTERED AND FORWARDED TO EXAMINER

STPP Information on status: patent application and granting procedure in general

Free format text: FINAL REJECTION MAILED

STPP Information on status: patent application and granting procedure in general

Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION

STPP Information on status: patent application and granting procedure in general

Free format text: NON FINAL ACTION MAILED

STPP Information on status: patent application and granting procedure in general

Free format text: RESPONSE TO NON-FINAL OFFICE ACTION ENTERED AND FORWARDED TO EXAMINER

STPP Information on status: patent application and granting procedure in general

Free format text: FINAL REJECTION MAILED

STPP Information on status: patent application and granting procedure in general

Free format text: ADVISORY ACTION MAILED

STCV Information on status: appeal procedure

Free format text: NOTICE OF APPEAL FILED

STCV Information on status: appeal procedure

Free format text: APPEAL BRIEF (OR SUPPLEMENTAL BRIEF) ENTERED AND FORWARDED TO EXAMINER

STPP Information on status: patent application and granting procedure in general

Free format text: AWAITING TC RESP., ISSUE FEE NOT PAID

STPP Information on status: patent application and granting procedure in general

Free format text: NOTICE OF ALLOWANCE MAILED -- APPLICATION RECEIVED IN OFFICE OF PUBLICATIONS

STPP Information on status: patent application and granting procedure in general

Free format text: PUBLICATIONS -- ISSUE FEE PAYMENT RECEIVED

STCF Information on status: patent grant

Free format text: PATENTED CASE